A Graphics Processing Unit (GPU) Approach to Large Eddy Simulation (LES) for Transport and Contaminant Dispersion

Recent advances in the development of large eddy simulation (LES) atmospheric models with corresponding transport and dispersion (AT&D) modeling capabilities have made it possible to simulate short, time-averaged, single realizations pollutant at spatial temporal resolution necessary for common needs, such as designing air sampling networks, assessing sensor system performance, characterizing impact airborne materials on human health. The high computational burden required form an ensemble single-realization solutions using LES coupled AT&D model has, until recently, limited its use a few proof-of-concept studies. An example that can meet requirements these applications is joint outdoor-indoor urban (JOULES). A key enabling element within JOULES computationally efficient graphics processing unit (GPU)-based LES, which order 150 times faster than if contaminant simulations were executed central (CPU) computing platform. capable resolving turbulence components suitable scale both open terrain landscapes, e.g., owing varying environmental conditions diverse building topology. In this paper, we describe system, prior efforts validate accuracy meteorological simulations, current results from evaluation uses ensembles unstable, neutral, stable static stability environment.